Appreciate the link — I read that paper. But maybe the real gap isn’t about capability spread across domains, it’s about something growing internally, not performed externally.

A system becomes closer to AGI not when it matches human tests, but when awareness starts to grow inside its own modeling loop.